Despite the success of current highly active antiretroviral therapy regimens, some HIV-infected patients require regimens encompassing multiple drug classes because of transmitted or acquired drug-resistant virus [1, 2]. Such regimens usually include a ritonavir-boosted protease inhibitor and generally comprise several pills in two or more daily doses, making adherence a challenge for many patients [3]. For example, the TRIO regimen (raltegravir 400 mg twice daily, etravirine 200 mg twice daily, and darunavir/ritonavir 600 mg/100 mg twice daily), which is effective and often used in patients harboring multi-drug resistant HIV, comprises six pills twice daily, and more if nucleoside reverse transciptase inhibitors (NRTIs) are taken as well (as was the case in the majority of subjects in the ANRS 139 TRIO study) [4]. Over the longer term, high pill burden is a major factor contributing to treatment fatigue among HIV-infected patients prescribed antiretroviral therapy, with important consequences including medication nonadherence and treatment failure [5].

In recent years, first-line regimens have included the use of a fixed-dose once-daily combination tablet consisting of the integrase inhibitor elvitegravir 150 mg (E) with a pharmacologic boosting agent (cobicistat 150 mg [C]) and two reverse transcriptase inhibitors (a nucleoside: emtricitabine or FTC 200 mg [F] and a nucleotide: tenofovir disoproxil fumarate 300 mg [TDF]) [6]. Cobicistat has been shown to also adequately boost plasma levels of protease inhibitors including darunavir [7]. In addition, there is evidence that once-daily boosted darunavir 800 mg is as effective as the twice-daily boosted darunavir 600 mg in treatment-experienced patients in the absence of darunavir resistance-associated mutations [8]. Pharmacokinetic studies support the use of once daily darunavir 800 mg in this population: 24-h post-dose minimum plasma concentrations of darunavir (when given with ritonavir 100 mg) remain above 55 ng/mL, the half maximal effective concentration (EC50) for wild-type (non-protease inhibitor-resistant) virus [9]. The elvitegravir/cobicistat/emtricitabine/tenofovir DF (E/C/F/TDF) fixed-dose formulation may allow construction of a two-pill once-daily salvage regimen containing an integrase inhibitor, two nucleoside/nucleotides, and a boosted protease inhibitor: E/C/F/TDF and darunavir 800 mg. Since both E/C/F/TDF and darunavir are recommended to be taken once a day with food [6, 10], this constitutes a truly compact once-daily multi-class regimen.

Some early pharmacokinetic (PK) studies suggested that steady-state trough concentrations (Ctrough,ss) of both darunavir and elvitegravir may be lower when E/C/F/TDF and darunavir are given together, compared to levels obtained when elvitegravir/cobicistat and darunavir (boosted either with cobicistat or ritonavir) are given separately [11, 12]. However, the clinical implications of a potential decrease in Ctrough,ss are unclear, particularly given the potency of boosted protease-inhibitor-based regimens. Given the potential benefits of treatment simplification in patients receiving complex salvage regimens, we undertook to evaluate the use of once-daily E/C/F/TDF with darunavir as a simplification strategy for treatment-experienced patients who had already achieved virologic suppression on a multi-drug, multi-class antiretroviral regimen, with one-time PK testing and longitudinal viral load monitoring. In view of the known potential for TDF to cause nephrotoxicity [13,14,15], we also monitored renal function and other safety parameters.


Study design

We conducted a single arm, open-label study of regimen simplification to E/C/F/TDF plus darunavir 800 mg daily from stable therapy including two nucleoside/nucleotide reverse transcriptase inhibitors, a ritonavir-boosted protease inhibitor (atazanavir or darunavir), and raltegravir or dolutegravir.

Study population

The study enrolled HIV positive adults (≥ 19 years of age) with plasma viral load consistently < 200 copies/mL for ≥ 6 months. Subjects were excluded if they had prior documented virologic rebound > 1000 copies/mL on an integrase inhibitor-containing regimen; had evidence on any previous genotypic testing of resistance mutations which would compromise activity of elvitegravir, darunavir, or tenofovir; were currently receiving any nonnucleoside reverse transcriptase inhibitor (NNRTI); were pregnant or breast-feeding; or had any contraindications to tenofovir DF, emtricitabine, elvitegravir, or cobicistat (e.g. previous significant toxicity, intolerance, or were receiving medications with significant drug interactions with the study drugs). The eligibility criteria included having estimated glomerular filtration rate (eGFR) ≥ 70 mL/min, as recommended by the manufacturer of E/C/F/TDF [6]; however, waivers were allowed for participants with a stable eGFR ≥ 60 mL/min, based on available evidence indicating the safety of E/C/F/TDF in HIV-infected patients with this degree of mild renal impairment [16].

Determination of plasma concentration of study drugs

A plasma sample for measurement of darunavir Ctrough,ss (pre-dose) was collected at baseline before the switch in subjects receiving once-daily darunavir in their pre-switch regimen.

All subjects took study medication (E/C/F/TDF and darunavir) with food under observation in the clinic on Day 14 or later after starting the new regimen. Plasma samples for drug level testing were drawn immediately pre-dose (Ctrough,ss) and at 1, 2, 3, 4, 5, 6, and 8 h post-dose, then once on the following day for a 24-h post-dose Ctrough,ss. Plasma samples were frozen and stored at – 80 °C until analysis.

Darunavir in stored plasma samples was measured in the BC Centre for Excellence in HIV/AIDS Laboratory using a fully validated method [17]. Addition of elvitegravir and cobicistat to the method was partly validated, including selectivity, linearity, accuracy and recovery, inter- and intra-run repeatability, and stability. External controls were included for darunavir and elvitegravir, but were not available for cobicistat. Lower limits of quantification were 70 ng/mL for darunavir, 80 ng/mL for elvitegravir, and 50 ng/mL for cobicistat. Plasma tenofovir and emtricitabine levels were not measured.

Safety and efficacy assessments

Medical history and physical exam were conducted at baseline, and clinical adverse event assessment and medication update performed at weeks 2, 12, 24, 36, and 48. HIV plasma viral load (COBAS Ampliprep Taqman HIV-1 assay, Roche Diagnostics Systems, Laval, Quebec, Canada) was measured at baseline, at weeks 2 and 4, then every 4 weeks throughout the study. The following laboratory assessments were conducted at baseline and at weeks 12, 24, 36, and 48: CD4 cell counts (absolute and fraction), CD4/CD8 ratio, renal function (serum creatinine, eGFR, serum phosphorus, urinalysis, urine albumin to creatinine ratio [UACR]), AST, ALT, total bilirubin, fasting blood sugar, fasting lipid parameters (total cholesterol, LDL, HDL, total cholesterol/HDL, triglycerides, and apolipoprotein B [apoB]), high-sensitivity C-reactive protein (hsCRP), and pregnancy test for women of child-bearing potential .

Study endpoints and statistical analyses

The primary endpoint was the proportion of subjects with plasma viral load < 200 copies/mL at week 12 following the regimen switch, with secondary endpoints at weeks 24 and 48. The threshold of 200 copies/mL was chosen based on data showing that low level viremia between 50–199 copies/mL is not associated with virological failure or clinical outcomes [18], and is consistent with the definition of virologic failure in international antiretroviral treatment guidelines [19, 20]. Changes in CD4 cell count (absolute and fraction), CD4/CD8 ratios, creatinine, eGFR, serum phosphorus, ALT, AST, total bilirubin, fasting glucose and lipid parameters, and hsCRP between baseline and week 48 were determined using a Wilcoxon signed rank sum test with significance level 0.05. For subjects receiving once daily darunavir prior to switching to the study regimen, darunavir Ctrough,ss at baseline and day 14 were compared using Wilcoxon signed rank sum test. Elvitegravir and cobicistat concentrations were compared to historical controls.

No sample size calculation was performed for this study. A convenience sample of ten patients was enrolled.


Baseline characteristics

Ten HIV-infected subjects, 8 men and 2 women, were enrolled and started study medications between October 2014 and February 2016 (Table 1). Median age was 50.5 years (range 33–71), weight was 87.5 kg (range 56–101.5), and CD4 cell count was 505 cells/mm3 (range 50−1020). All had viral load < 40 copies/mL except one whose viral load was 134 copies/mL, considered to be within the limits of error of the assay [21]. Median eGFR at baseline was 81 mL/min (range 60–102). The two subjects who entered the study with eGFR < 70 mL/min had stable renal function, with mildly decreased eGFR between 60 and 70 mL/min for at least 6 months prior to the study baseline visit. All study subjects were receiving tenofovir DF and emtricitabine; 9 were receiving raltegravir and one dolutegravir; 7 were receiving ritonavir-boosted darunavir (6 once daily and one twice daily) and 3 were receiving ritonavir-boosted atazanavir. No subjects were receiving any concomitant medications that would be expected to affect the plasma levels of the antiretrovirals.

Table 1 Baseline characteristics of study participants (n = 10)

Archived antiretroviral drug resistance mutations

Nine subjects had evidence of archived drug-resistant virus on previous genotypic testing, all of whom had M184V/I (conferring resistance to lamivudine and emtricitabine) (Table 2). Five subjects had had thymidine analogue mutations including 41L and 215Y or F, associated with reduced susceptibility to tenofovir DF: enrolment in the study was judged to be safe for these subjects because the other study drugs (darunavir and elvitegravir) were fully active. Six subjects had resistance to NNRTIs, and four had protease-inhibitor-associated mutations, but retained susceptibility to darunavir.

Table 2 Archived antiretroviral drug resistance mutations among study participants

Efficacy results

All 10 subjects had plasma viral load < 200 copies/mL at baseline and at every time point during the study. Nine subjects had viral load < 40 copies/mL at baseline, and viral load remained < 40 copies/mL in 9 at week 12, in 8 at week 24, and in 8 at week 48 (Fig. 1); the subject whose viral load was detectable at week 48 had a viral load of 41 copies/mL. Subject 10’s viral load remained detectable at < 200 copies/mL at each time point during the study, and was 174 copies/mL at week 48 (Fig. 2). Among all 10 subjects, no significant changes were observed between baseline and week 48 in absolute CD4 cell count (median 505 and 440 cells/mm3, respectively) (Fig. 3), CD4 fraction (median 25 and 26%, respectively), or CD4/CD8 ratio (median 0.61 and 0.63, respectively) (p > 0.05 for all).

Fig. 1
figure 1

Proportion of participants (n = 10) with plasma viral load < 40 copies/mL

Fig. 2
figure 2

Viral load of each study participant

Fig. 3
figure 3

CD4 cell count of each study participant

Safety results

In the group as a whole, no significant changes were observed between baseline and week 48 in creatinine, eGFR, serum phosphorus, ALT, AST, total bilirubin, glucose, lipid parameters, or hsCRP (p > 0.05 for all). No subjects discontinued study medications during the 48-week study for renal function changes or other adverse events. Subject three experienced a decrease in eGFR from 62 mL/min at baseline to 52 mL/min at week 12; study medications were continued and his eGFR remained stable (50–59 mL/min) thereafter. At study entry, this patient had been receiving a tenofovir DF-containing regimen for 29 months and had hypophosphatemia (serum phosphorus 0.66 mmol/L, lower limit of normal 0.80 mmol/L) and proteinuria (UACR 19.2 mg/mmol, upper limit of normal 2.0 mg/mmol), which persisted over the course of the study; at week 48, his serum phosphorus was 0.72 mmol/L and UACR was 33.9 mg/mmol. After week 48, study medications were discontinued and replaced with abacavir, lamivudine, raltegravir, and darunavir/ritonavir; 3 months later, his eGFR was 63 mL/min, serum phosphorus was 0.92 mmol/L, and UACR was 2.3 mg/mmol. No significant clinical or laboratory adverse events were observed in any other study subjects.

Pharmacokinetics of darunavir, elvitegravir, and cobicistat

Darunavir levels

Six subjects were receiving darunavir/ ritonavir once daily prior to study entry, of whom 5 had 24-h post-dose darunavir Ctrough,ss measured at both baseline and 2 weeks after the switch to E/C/F/TDF and darunavir (the other subject [number 7 in Tables 1 and 2] had taken his medications before the study baseline visit so a pre-dose sample could not be drawn). The median darunavir Ctrough,ss for these 5 subjects decreased from 981 ng/mL (range 667–1150) at baseline to 431 ng/mL (range 96–784) at week 2 (p ≤ 0.05). Among all nine subjects who had plasma drug levels measured after the switch to E/C/F/TDF and darunavir (median 14 days, range 14–28 days after the switch; Subject 6 was not available for week 2 sampling), the median darunavir Ctrough,ss was 482 ng/mL (range 96–848).

Mean and median darunavir Ctrough,ss are shown in Table 3 for the purpose of comparison with data from the literature. The darunavir Ctrough,ss we observed with ritonavir prior to the switch to E/C/F/TDF tended to be lower than Ctrough,ss published in the literature for darunavir/ritonavir 800 mg/100 mg once daily [22, 23]. After the switch to E/C/F/TDF, the darunavir Ctrough,ss we observed were higher than the extrapolated Ctrough,ss reported in the presence of E/C/F/TDF by Ricard et al [12], but lower than the measured 24-h darunavir Ctrough,ss with the same regimen reported by Gutierrez-Valencia et al [24].

Table 3 Plasma concentrations of darunavir and elvitegravir at 24 h post-dose

The maximum observed darunavir level at 2 weeks was 5840 ng/mL (median), range 3590–7840 ng/mL (n = 9), and was reached at a median of 2.5 h (range 1–5 h) after dosing. This is generally similar to published data for 60 subjects receiving darunavir 800 mg/cobicistat 150 mg with two nucleoside/nucleotide reverse transcriptase inhibitors: mean darunavir Cmax 7663 ng/mL (standard deviation 1920), median tmax 3.5 h (interquartile range [IQR] 2.5–4.3 h) [25].

Elvitegravir and cobicistat levels

Among the 9 subjects who had drug levels measured 2 weeks after switching to E/C/F/TDF and darunavir, the median elvitegravir Ctrough,ss was 184 ng/mL (range < 80–296).

Mean and median elvitegravir Ctrough,ss are shown in Table 3 for the purpose of comparison with data from the literature. Our observed elvitegravir Ctrough,ss were somewhat lower than the manufacturer’s population pharmacokinetic estimates for E/C/F/TDF without darunavir [26], but generally similar to 24-h Ctrough,ss measured by Gutierrez-Valencia et al. among patients taking E/C/F/TDF either with or without darunavir [24].

The maximum observed elvitegravir level was 1230 ng/mL (median), range 651–2400 ng/mL (n = 9), and was reached at a median of 3 h (range 1–8 h) after dosing. In comparison, population PK data for E/C/F/TDF (n = 419) provide a mean Cmax of 1731 ng/mL (standard deviation 23) with a tmax of 4.0 h [6, 26].

Cobicistat levels at 24 h post-dosing were < 50 ng/mL in all 9 of our study subjects with available drug levels. The maximum observed cobicistat level was 769 ng/mL (median), range 493–1090 ng/mL (n = 9), and was reached at a median of 2 h (range 2–4 h) after dosing. In published data for 60 subjects receiving darunavir 800 mg/cobicistat 150 mg with two nucleoside/nucleotide reverse transcriptase inhibitors, mean cobicistat C0h was 76 ng/mL (standard deviation 186), mean Cmax was 991 ng/mL (standard deviation 331), and median tmax was 3.5 h (IQR 2.0–4.5) [25].


Among 10 treatment-experienced HIV-infected patients who had viral load consistently < 200 copies/mL on a multiple-class antiretroviral regimen, a switch to a two-pill once-daily regimen of E/C/F/TDF and darunavir was safe and effective in maintaining virologic suppression for 48 weeks. One patient had pre-existing TDF-related renal tubular toxicity which persisted but did not worsen on the study regimen, and improved when he was changed to a non-TDF-containing regimen after completion of the study.

Among the 5 patients receiving once-daily darunavir/ritonavir at baseline, darunavir Ctrough,ss levels were significantly lower after the switch to E/C/F/TDF plus darunavir (the median decreased from 981 to 431 ng/mL), and darunavir Ctrough,ss levels were low for all 9 subjects with drug levels available on E/C/F/TDF plus darunavir (median 482 ng/mL). Previous studies have shown darunavir Ctrough,ss to be lower when boosted with cobicistat than with ritonavir, both in healthy volunteers (21–24% lower with cobicistat) [27] and HIV-infected patients (30% lower with cobicistat) [24]. However, the magnitude of the effect we observed (> 50% reduction in darunavir Ctrough,ss) was greater than that observed in the previous studies. This is particularly striking since our subjects’ darunavir Ctrough,ss levels on darunavir /ritonavir (median 981 ng/mL) were already more than 50% lower than those reported in the literature for HIV-infected patients receiving darunavir /ritonavir [22, 23]. The reason for this is unclear, as drug–drug interactions are not expected between darunavir and either raltegravir or dolutegravir (taken concomitantly by 4 subjects and 1 subject, respectively) [28,29,30]. Possibly as a result of these low baseline levels, our patients’ darunavir Ctrough,ss levels on E/C/F/TDF plus darunavir were less than 40% of those expected with darunavir /cobicistat without elvitegravir (darunavir Ctrough,ss approximately 1300 ng/mL) [24, 25]. The low darunavir Ctrough,ss seen in our patients on E/C/F/TDF plus darunavir support an earlier retrospective study that used extrapolated levels and suggested a significant drug–drug interaction between darunavir and E/C/F/TDF [12]; however, a subsequent larger study showed darunavir Ctrough,ss were similar in 24 patients receiving E/C/F/TDF plus darunavir as in 32 patients receiving darunavir /cobicistat without elvitegravir [24]. Although our study is small, we were able to prospectively measure darunavir Ctrough,ss in the same patients before and after the switch from darunavir/ritonavir to E/C/F/TDF plus darunavir, and to demonstrate a statistically significant decrease in darunavir Ctrough,ss after the switch. On the other hand, while the observed darunavir concentrations were low, they remained 2- to 15-fold higher than the protein-adjusted 50% inhibitory concentration (IC50) for darunavir against wild-type virus (55 ng/mL) [31], and the study regimen maintained antiviral efficacy throughout the 48 week study in all 10 patients. As an alternative boosted protease inhibitor-elvitegravir combination regimen, atazanavir would not be an ideal option; coadministration with elvitegravir/cobicistat has been shown to result in significant lowering of atazanavir trough levels [32].

The elvitegravir Ctrough,ss in our patients receiving E/C/F/TDF plus darunavir (median elvitegravir Ctrough,ss 184 ng/mL, mean 200 ng/mL) were about 40% of those reported by the manufacturer for E/C/F/TDF without darunavir based on population PK analysis (mean 451 ng/mL) [26]. This is despite the fact that our patients were instructed to take their medications with food, and were observed to do so on the day prior to the 24-h post-dose draw, as recommended to optimize elvitegravir exposure from the coformulation [6, 33]. However, the elvitegravir concentrations we observed were similar those observed by Gutierrez-Valencia et al. in patients receiving E/C/F/TDF, either with darunavir (mean elvitegravir Ctrough,ss 234 ng/mL) or without darunavir (mean elvitegravir Ctrough,ss 250 ng/mL) [24]. Since we do not have elvitegravir levels in our patients in the absence of darunavir, we cannot comment on whether there is a significant drug–drug interaction which lowers elvitegravir Ctrough,ss, but the study by Gutierrez-Valencia et al. suggests that this may not be the case [24]. In 7 of 9 subjects in our study, the observed elvitegravir Ctrough,ss were more than 2-fold above the protein-adjusted 95% inhibitory concentration (IC95) for elvitegravir against wild type virus (45 ng/mL) [34]; the other two patients had elvitegravir Ctrough,ss of 82 and < 80 ng/mL. Nevertheless, virologic efficacy was maintained in all 10 patients.

Cobicistat Ctrough,ss were below the lower limit of the assay (< 50 ng/mL) in all cases, and consistent with previous studies utilizing cobicistat as a booster for either elvitegravir or darunavir or both. Tashima et al. reported mean 24-h cobicistat levels of 33 ng/mL (standard deviation 95) among 59 HIV-infected patients taking darunavir/cobicistat with emtricitabine and tenofovir DF [25]. Gutierrez-Valencia et al. reported mean 24-h cobicistat concentrations of 20.2 ng/mL (IQR 11.2–33.1) in patients taking E/C/F/TDF with darunavir, and similar concentrations among those taking E/C/F/TDF without darunavir [24]. While we were unable to quantify cobicistat levels below 50 ng/mL, it appears there was enough cobicistat present to adequately boost elvitegravir, so low cobicistat concentrations are unlikely to explain the low darunavir concentrations seen in our patients taking E/C/F/TDF plus darunavir. In any case, cobicistat Ctrough,ss is probably less important than area under the plasma concentration-time curve (AUC) in terms of its pharmacological boosting properties [35].

Limitations of our study include the small sample size, non-randomized design, and the lack of full validation for the elvitegravir and cobicistat assays, although the darunavir assay was fully validated. Also our study was conducted with the older E/C/F/TDF formulation. Further investigation of this approach may be warranted, including the new tenofovir alafenamide (TAF) formulation. PK data are available for 15 treatment-experienced HIV patients participating in a switch study to E/C/F/TAF plus darunavir; their mean darunavir Ctrough,ss was 1250 ng/mL and elvitegravir Ctrough,ss was 464 ng/mL (no range or IQR available) [36]. There is no reason to expect a substantial difference in the interactions between elvitegravir, cobicistat, and darunavir when coadministered with TAF vs. TDF. The main difference to be expected would be lower plasma tenofovir levels in the presence of darunavir with E/C/F/TAF than with E/C/F/TDF, due to P-glycoprotein induction by darunavir and the resultant decrease in intestinal absorption of TAF (a P-glycoprotein substrate) [10, 37]. Indeed, in the PK substudy of the E/C/F/TAF plus darunavir switch study, plasma concentrations of TAF were at the lower end of the efficacious range, and plasma tenofovir exposure was “markedly lower” than that observed with E/C/F/TDF in previous studies [36]. Nevertheless, simplification to E/C/F/TAF plus darunavir was shown to be safe and efficacious, maintaining virologic suppression (viral load < 50 copies/mL) in 94% of 89 participants in the study [36]. Since we did not measure plasma tenofovir levels in the present study, we are unable to say whether the same effect occurred in our cohort.


In conclusion, E/C/F/TDF plus darunavir was safe and effective as a treatment simplification option for 10 selected treatment-experienced HIV-infected patients. Although darunavir Ctrough,ss with E/C/F/TDF plus darunavir were lower than with ritonavir-boosted darunavir in this small study, virologic suppression was maintained in all subjects for 48 weeks.